Chorus CLI
Chorus convenes 2–4 AI reviewers from different CLIs to check a change before release, and plugs into agents as an MCP server.
- 533 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Code is sent to several external AI services.
- In Full mode reviewers write outside the working directory and reach the network.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/chorusDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Code is sent to several external AI services.
- In Full mode reviewers write outside the working directory and reach the network.
This is third-party code. Review the repository files before installing.
What it does
Chorus runs several AI tools in parallel: one writes or proposes a change, the others review it. Approval comes when the reviewers agree. It works through already installed Claude Code, Codex CLI, Gemini CLI, OpenCode and Kimi CLI, or through an OpenRouter key. `chorus init` finds such tools and registers Chorus as an MCP server in Claude Code, Codex, Gemini CLI, Cursor, Windsurf, Kimi and OpenCode. It includes review templates (`code-review`, `bug-diagnose`, `architect-review`, `red-green`, `review-only`), reviewer personas and a web panel on `localhost:5050`.
Who it is for. Developers writing code with AI who want a second opinion from models of another vendor before merging.
Good fit when
- An agent rewrote 200 lines of a critical path and claims the behaviour is unchanged.
- Before merging a big pull request a quick review from three models is needed.
- An architecture decision (queue or polling, sync or async) should be examined from several angles.
- One model writes tests blind to the code and another writes code to pass them.
- A flaky bug is hunted by showing the suspect code to several reviewers.
Not a fit when
- Only one AI tool is available, and review loses its point without different vendors.
- Code cannot be sent to any external AI service.
- An instant check is needed, while several models running in parallel take minutes.
Example request
Use chorus to review the staged diff against main and flag race conditions and missing tests.Limitations
Node 20+ and at least one CLI (Claude Code, Codex CLI, Gemini CLI, OpenCode, Kimi CLI) or an OpenRouter key are required. Usage counts against your subscription quotas or paid keys. In `Full` mode a reviewer writes outside its working directory and reaches the network. For non-interactive `codex exec` the README requires the `--dangerously-bypass-approvals-and-sandbox` flag. Chorus sends fixed telemetry once a day without chat content. The last commit dates to August 2026.
How to disable. Stop the daemon with `chorus stop`, remove the package with `npm uninstall -g chorus-codes` and delete the `chorus` entry from MCP client configs. Data lives in `~/.chorus`.
FAQ
What does a review cost?
There is no markup. Usage counts against your subscription quotas, and with API keys the README puts a typical run at 0.30–1.50 dollars.
Which access modes do reviewers have?
`Strict` reads code without writes or network, `Workspace` (default) writes inside the working directory, `Full` writes anywhere and reaches the network.
Which MCP tools exist?
Nine: `create_chat`, `wait_for_chat`, `get_chat_status`, `cancel_chat`, `resume_chat`, `list_templates`, `list_personas`, `invoke_persona`, `list_blocked`.
How do I create my own template?
Put a YAML file in `~/.chorus/templates/` or copy a built-in template and change reviewer slots and quorum.
How do I collect data for a bug report?
`chorus diagnose` prints a redacted diagnostic block.
Discussion
No questions or reviews yet.
Related
CLI toolsA skill and CLI for code review: a team of reviewer agents inspects changes independently, debates findings and produces one report
A code-review CLI and plugin for agents: reads a git diff and writes line-level comments with context
A Claude Code plugin that hands a task to a local no_human and reviews a branch with a reviewer that never saw the coder's session
A CLI that runs Claude Code, Codex, Gemini CLI and other agents in parallel on one task and collects their answers side by side
A Claude Code plugin: Claude does the task, then a stop hook starts an independent Codex review and Claude fixes the findings in the same session
Kimi Bridge for Claude Code
kimi-bridge
An MCP server that turns a local Kimi Code CLI into a second agent inside Claude Code, with a dedicated BSL review tool for 1C
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Chorus (this card) | no | 533 | |
| Claude-Codex Review Skill | no | 5 | a skill for a Claude and Codex pair; Chorus runs up to four models and manages templates with quorum |
| open-code-review | no | 49,484 | reviews code against a set of rules; Chorus combines opinions of several models from different vendors |
README in short
The README explains cross-vendor review, templates, personas, access modes and the fixed telemetry payload.